Cannabinoid-mediated antinociception is enhanced in rat osteoarthritic knees

Authors

  • Niklas Schuelert,

    1. University of Calgary, Calgary, Alberta, Canada
    Search for more papers by this author
  • Jason J. McDougall

    Corresponding author
    1. University of Calgary, Calgary, Alberta, Canada
    • Department of Physiology and Biophysics, University of Calgary, 3330 Hospital Drive NW, Calgary, Alberta T2N 4N1, Canada
    Search for more papers by this author
    • Dr. McDougall is an Alberta Heritage Foundation for Medical Research Senior Scholar and is recipient of an Arthritis Society of Canada Investigator award.


Abstract

Objective

To determine whether local administration of the cannabinoid 1 (CB1) receptor agonist arachidonyl-2-chloroethylamide (ACEA) can modulate joint nociception in control rat knee joints and in experimental osteoarthritis (OA).

Methods

OA was induced in male Wistar rats by intraarticular injection of 3 mg of sodium mono-iodoacetate, with a recovery period of 14 days. Electrophysiologic recordings were made of knee joint primary afferent nerve fibers in response to normal rotation and noxious hyperrotation of the joint both before and after close intraarterial injection of different doses of ACEA.

Results

Local application of the CB1 agonist significantly reduced the firing rate of afferent nerve fibers by up to 50% in control knee joints (n = 19) and up to 62% in OA knee joints (n = 29; P < 0.01). Coadministration of the CB1 receptor antagonist AM251 or the transient receptor potential vanilloid 1 (TRPV-1) ion channel antagonist SB366791 significantly reduced the desensitizing effect of ACEA. The CB1 receptor antagonist AM251 by itself had no effect in the control joint but significantly increased the firing rate of afferent nerve fibers in the OA joint.

Conclusion

These findings indicate that activation of peripheral CB1 receptors reduces the mechanosensitivity of afferent nerve fibers in control and OA knee joints. Blockade of either the CB1 receptor or the TRPV-1 channel significantly reduced the efficacy of ACEA, which suggests that both receptors are involved in cannabinoid-mediated antinociception. The increased nerve activity observed following CB1 receptor antagonism suggests a tonic release of endocannabinoids during OA. As such, peripheral CB1 receptors may be important targets in controlling OA pain.

Osteoarthritis (OA) is the most common form of arthritis, affecting >10 million people worldwide. OA is characterized by extensive remodeling of subchondral bone and permanent destruction of articular cartilage, which lead to structural and functional degradation of affected synovial joints. OA is often associated with symptoms of pain that typically worsen with weight bearing and activity. Currently no disease-modifying drugs are available for OA; therefore, treatment is primarily restricted to analgesics, which often have limited efficacy and hazardous side effects.

An established animal model of OA pain involves the intraarticular injection of the glycolysis inhibitor sodium mono-iodoacetate (MIA), which disrupts cartilage metabolism, thus leading to chondrocyte death and subchondral bone lesions consistent with the pathologic changes seen in the human condition (1, 2). Electrophysiologic and behavioral studies have shown that joint nociceptors are sensitized in the MIA model, leading to the generation of joint pain (3, 4). One of the main mechanisms responsible for the generation of joint pain is the activation of nociceptors located on the terminal branches of joint type III (Aδ fiber) and type IV (C fiber) primary afferent nerve fibers (5, 6). These afferent nerve fibers show increased activity when a noxious stimulus is applied to the innervated tissue, leading to the activation of central pathways and the experience of pain (6, 7). One approach to alleviating joint pain is the inhibition of joint nociceptor activity, which would reduce noxious sensory input to the central nervous system.

Agents that have shown promise in the treatment of chronic pain are the cannabinoids, which are a group of alkaloids derived from the hemp plant Cannabis sativa. Cannabinoids are also synthesized in mammalian tissues (so-called endocannabinoids), where they play a major role in multiple physiologic processes. The discovery of endocannabinoids has opened up an exciting new approach to pain management because their antinociceptive effects have been reported in several pain models (8–10); however, the exact mechanism of these analgesic effects requires further elucidation. Cannabinoids are known to bind to 2 receptor subtypes, cannabinoid 1 (CB1) and CB2 (11, 12). CB1 receptors are present at all levels of the pain pathway, including the primary afferent fibers, spinal cord, and supraspinal sites (10, 13, 14), whereas CB2 receptors occur mainly on immune cells (14–16), where they modulate the release of inflammatory mediators (17).

Electrophysiologic and behavioral studies have provided convincing evidence that cannabinoids suppress nociceptive transmission at the peripheral and central levels of the pain pathway. Cannabinoids alter neurotransmission through CB1 receptors by inhibiting Ca2+ channel (18, 19) and adenylate cyclase (20) activity and by activation of K+ channels and MAP kinase (21). CB1 receptors are synthesized in dorsal root ganglia (22–24) and are transported to peripheral nerve terminals, where their activation can inhibit tissue nociception. It has been shown that CB1 agonists, applied locally or systemically, can reduce hyperalgesia after the induction of acute or chronic inflammation (10, 25, 26). Whether selective activation of cannabinoid receptors reduces mechanosensitivity of afferent nerve fibers in a model of OA in which inflammation is considered to be a secondary component has not been tested so far.

In addition to being agonist ligands at cannabinoid receptors, several cannabinoids are also activators of the transient receptor potential vanilloid channel 1 (TRPV-1) (27–29). TRPV-1 is a nonselective cation channel with 6 transmembrane-spanning domains that is typically activated by noxious heat (>43°C), low pH, and naturally occurring vanilloids such as capsaicin and resiniferatoxin (30, 31). TRPV-1 is expressed on nociceptive afferent neurons throughout the periphery and has been shown to play a critical role in the induction of thermal hyperalgesia in inflammatory pain models (32–34). The endocannabinoid anandamide and the selective CB1 receptor agonist arachidonyl-2-chloroethylamide (ACEA) have been shown to activate TRPV-1 channels, which leads to neuronal inward currents and vasodilatation (27–29). Thus, certain cannabinoids may act as dual cannabinoid–vanilloid regulators, particularly under inflammatory hyperalgesia conditions, where increased coexpression of CB1 receptors with TRPV-1 channels has been observed (35).

In this study, we examined whether local application of the CB1 receptor agonist ACEA could reduce nociceptive activity of afferent nerve fibers in saline-injected control and OA knee joints of rats. Furthermore, we tested the specificity of this agonist by coadministering the selective CB1 receptor antagonist AM251 and the TRPV-1 receptor antagonist SB366791.

MATERIALS AND METHODS

Animals.

Experiments were performed on 95 male Wistar rats (250–450 gm), which were housed in cages with free access to water and rodent food and maintained at room temperature (22°C) under a 12-hour light/12-hour dark cycle. The animal handling and surgical procedures outlined in this study were performed in accordance with the Canadian Council for Animal Care guidelines for the care and use of experimental animals.

Surgical procedures.

Rats were deeply anesthetized using urethane (25% stock solution; 2 gm/kg intraperitoneally [IP]) before any surgical procedures. The depth of the anesthesia was confirmed by the absence of the hind paw withdrawal reflex. Core body temperature was measured by a rectally inserted thermometer and was maintained at 37°C by a thermostatically controlled heating blanket. The trachea was cannulated, and the cannula was connected to a Harvard rodent respiratory pump for artificial ventilation with 100% O2 (stroke volume of 2.5 ml, breath frequency set at 60 breaths/minute; Harvard Apparatus, Holliston, MA). The left carotid artery was then cannulated with a fine-bore catheter (0.5-mm inner diameter [ID], 1.00-mm outer diameter [OD]; Portex, Kent, UK) containing heparinized saline (100 units/ml). The cannula was connected to a pressure transducer to allow for continuous blood pressure measurement with a blood pressure monitor (BP-1; World Precision Instruments, Sarasota, FL). The left jugular vein was also cannulated (fine-bore tubing, 0.40-mm ID, 0.80-mm OD; Portex) and the muscle relaxant gallamine triethiodide (50 mg/kg) was injected once through the catheter to eliminate neural interference from the hind limb musculature.

A catheter was introduced into the right saphenous artery below the knee joint and advanced to a point just distal to the bifurcation with the medial articular artery to permit local close intraarterial injection of drugs to the knee joint. The central portion of the right femur was isolated, and a specialized clamp was fixed to the midshaft of the bone and attached to a stereotaxic frame to immobilize the proximal aspect of the rat hind limb. The right hind paw was then placed in a shoe-like holder that was connected to a force transducer and torque meter (MVD2510; HBM, Darmstadt, Germany) to standardize the amount of rotational force being applied to the knee joint. Finally, a longitudinal skin incision was made along the medial aspect of the hind limb, and the resulting skin flaps were secured to a metal O-ring to create a pouch that was filled with warm paraffin oil. This oil pool served to prevent tissue desiccation throughout the experiment.

Extracellular electrophysiologic recording.

The technique used for recording rat knee joint afferent nerve fibers has been described previously (6, 36). Briefly, the saphenous nerve was transected distally to the knee joint to eliminate sensory input from the foot and ankle region. The saphenous nerve was also isolated in the inguinal region and cut centrally to prevent the generation of spinally mediated reflexes. The saphenous nerve stump projecting centrally from the knee was placed on a small, black Perspex stage to facilitate the isolation of fine neurofilaments, which were dissected free from the nerve using fine watchmaker forceps. Nerve fibers were then placed over a platinum electrode to permit extracellular recordings. Afferent nerve fibers originating from the knee joint were identified by the elicitation of a response to gentle probing of the knee joint with a glass rod with a 1-mm tip.

The conduction velocity of the nerve fibers was determined by electrically stimulating their receptive field with a pair of bipolar silver wire electrodes (1 Hz, 100 msec pulse width, 3–10V); the distance between stimulating and recording electrodes was divided by the latency between stimulus artifact and evoked afferent impulse to obtain the conduction velocity. The electrical threshold of each fiber was determined by placing the bipolar silver electrode in the receptive field of the fiber and gradually increasing the voltage of stimulation in 0.5V steps until the pulse-elicited spike discharges from the afferent nerve. The mechanical threshold of individual units was also determined by slowly rotating the knee until nerve activity could be detected and noting the force required to elicit such a response on the torque meter. The mechanosensitivity of articular afferent nerve fibers was tested by recording nerve activity in response to outward non-noxious rotation and noxious hyperrotation of the knee joint. Non-noxious rotation is defined here as movement occurring within the normal working range of the joint, while noxious hyperrotation of the joint is defined here as movement occurring outside the normal working range of the knee without imparting overt tissue trauma. This level of joint rotation causes maximal activation of joint primary afferent nerve fibers, which would result in a painful sensation in an alert animal. The amount of force required to produce this noxious mechanical stimulus was between 20 and 40 mNm.

Three movement cycles, each consisting of a normal rotation and hyperrotation of the knee to discrete torque levels, were performed at the beginning of the experiment, and the mean afferent nerve fiber firing rate associated with these movements was the control baseline level, which was set at 100%. Each movement lasted 10 seconds, and the same level of rotation was repeated every 2 minutes until 15 minutes after drug administration. Once fiber activity returned to control levels, the next dose of drug was applied to the knee. In the control and OA groups, recordings were made before (control) and after close intraarterial injection of ACEA. In separate experiments, the CB1 receptor antagonist AM251 (10−8 moles; 0.1-ml bolus) or the TRPV-1 receptor antagonist SB366791 (500 μg/kg IP) was administered prior to the ACEA to confirm cannabinoid and/or TRPV-1 receptor involvement in joint mechanosensitivity. Antagonists were administered by close intraarterial injection immediately prior to each dose of ACEA. Neuronal activity was digitized using a data acquisition system (CED1401; Cambridge Electronic Design, Cambridge, UK) and stored on a microcomputer for offline analysis. The number of action potentials (APs) per movement was determined using Spike 2 software (Cambridge Electronic Design), and the percentage change in afferent nerve fiber activity was calculated.

Induction of OA.

Forty rats were deeply anesthetized with 2% isoflurane in 100% O2 (1 liter/minute) until the flexor withdrawal reflex was abolished. The skin overlying the right knee joint was shaved and swabbed with 100% ethanol, and the diameter was measured across the joint line in a mediolateral plane using a digital micrometer (Mitutoyo Instruments, Tokyo, Japan). A 27-gauge needle was introduced into the joint cavity through the patellar ligament, and 50 μl of 3 mg MIA in 0.9% saline was injected into the joint to induce OA-like lesions. Animals were allowed to recover for 14 days, during which time it has consistently been shown that severe end-stage OA gradually develops in this species (2, 37, 38). This was confirmed in our experiments, in which joint diameter was significantly increased at this time point. Measurements of OA knee joints were performed as described above, and the effect of ACEA, AM251, and SB366791 on joint mechanosensitivity was assessed.

Drugs and reagents.

In vitro binding assays have confirmed that ACEA has a much greater affinity for the CB1 receptor (mean ± SEM Ki = 1.4 ± 0.3 nM) compared with the CB2 receptor (Ki = 3.1 ± 1.0 μM) (39). Furthermore, potency and efficacy studies using the 35S-GTPγS binding assay in rat cerebellar membranes have confirmed that ACEA is a highly potent full agonist at the CB1 receptor (39). A number of in vitro and in vivo studies have confirmed that AM251 is a potent antagonist at the CB1 receptor (Ki = 7.49 nM), with a 300-fold selectivity over the CB2 receptor (for review, see ref.40). SB366791 is a competitive antagonist at the TRPV-1 receptor, where it has been shown to antagonize chemical agonists, noxious heat, and acid (41).

MIA, ACEA, AM251, and SB366791 were obtained from Tocris (Ellisville, MO); gallamine triethiodide, DMSO, Cremophor, and urethane were obtained from Sigma-Aldrich (Oakville, Ontario, Canada). All reagents were dissolved in vehicle solution (2% DMSO, 1% Cremophor, 0.9% saline), and aliquots of the drug were kept frozen (−20°C) in Eppendorf vials (Eppendorf, Madison, WI) until required. Gallamine triethiodide was made fresh on the day of experimentation and dissolved in 0.9% saline.

Statistical analysis.

Data were expressed as the mean ± SEM. The effect of drugs between animal groups was analyzed by two-way analysis of variance (ANOVA) with Bonferroni adjustment and by Student's unpaired t-test. Dose dependency and the time course of drugs were tested by one-way ANOVA, with individual points being compared with the control by a 1-sample t-test. The maximal efficacy (Emax) of ACEA was calculated from dose-response curves. All data passed a normality test; therefore, a Gaussian distribution can be assumed and parametric statistics were applied. P values less than 0.05 were considered significant.

RESULTS

Neuronal characteristics.

Between 1 and 3 afferent fibers were examined per animal, such that a total of 159 units (81 units in control knee joints, 78 units in OA knee joints) were recorded in this study. Compared with saline-injected control knees, the average firing rate of afferent nerve fibers in OA knee joints was significantly higher during normal rotation (saline-injected joints, mean ± SEM 43.8 ± 4.53 APs/10 seconds of movement; OA joints, 54.3 ± 4.91 APs/10 seconds of movement; [P < 0.01 by Student's unpaired t-test]) and during noxious hyperrotation (saline-injected joints, 76.6 ± 6.19 APs/10 seconds of movement; OA joints, 111.9 ± 6.77 APs/10 seconds of movement [P < 0.001]).

Furthermore, the mean ± SEM mechanical threshold required to initiate firing of afferent nerve fibers in control joints was 14.5 ± 0.73 mNm, while in OA knees, the threshold was significantly reduced to 11.9 ± 0.68 mNm (P < 0.01 by Student's unpaired t-test). The electrophysiologic characteristics of these units are summarized in Table 1. All units tested could be activated by local injection of KCl (0.4 mM; 0.1 ml) at the end of the experiment, confirming that administered reagents reached mechanosensory nerve endings throughout the experiment. Fibers responded to outward rotation of the knee and had a mechanical threshold that was within the normal working range of the knee joint, i.e., typically only low threshold units were recorded. Two fibers with a high mechanical threshold (i.e., responsive only to noxious movements) were also included in the study.

Table 1. Proportion and electrophysiologic characteristics of thinly myelinated type III and unmyelinated type IV fibers recorded in control and osteoarthritic (OA) knee joints
 Mechanical threshold, mNmElectrical threshold, VConduction velocity, meters/second% of fibers
  • *

    Basal firing frequency range was 0–195 action potentials (APs)/movement for normal movement and 4–233 APs/movement for noxious movement.

  • Basal firing frequency range was 3–175 APs/movement for normal movement and 25–346 APs/movement for noxious movement.

Control joints*    
 Type III7–345–122.6–18.648.2
 Type IV7–354.5–80.5–2.451.8
OA joints    
 Type III6–166–82.9–8.025.3
 Type IV3–255–80.6–2.474.7

Effect of ACEA in control knee joints.

With normal or noxious joint rotation, injection of ACEA induced a marked suppression of nociceptive activity in 54% of recorded fibers in the control joints. No change in the firing rate was seen in 38% of recorded fibers, while 8% of recorded fibers showed a maximum response to application of the lowest dose of ACEA (hyperresponding fibers) (Figure 1A). Application of vehicle had no significant effect on mechanosensitivity (P = 0.42 for non-noxious movement and P = 0.56 for noxious movement, by 1-sample t-test; n = 15 fibers). The desensitizing effect of ACEA was maximal 5 minutes after drug application for both types of mechanical stimuli (P < 0.01 by 1-sample t-test; n = 10–11 fibers) and lasted >30 minutes in most fibers. A specimen recording of the inhibitory effect of ACEA on afferent nerve activity is shown in Figures 1B and C. The desensitizing effect of ACEA was found to be dose-dependent across the dose range of 10−10 moles to 10−7 moles, for non-noxious movements (P < 0.05 by one-way ANOVA). The mean ± SEM Emax of ACEA during normal rotation was −50.0 ± 5.69%, and for noxious rotation the Emax was −36.4 ± 4.05%. Compared with the vehicle, the desensitizing effect of ACEA was statistically significant (P < 0.0001 by two-way ANOVA) (Figures 1D and E).

Figure 1.

Effect of arachidonyl-2-chloroethylamide (ACEA) on control knee joints. A, Proportion of afferent nerve fibers from control (saline-injected) knee joints that showed no change in activity, a dose-dependent decrease in activity, and hyperresponsiveness to local application of ACEA. B and C, Specimen recording of a single unit during normal and noxious rotation of a control knee joint before (B) and after (C) close intraarterial application of ACEA. ACEA significantly reduced the firing rate of joint afferent nerve fibers. D and E, Effect of ACEA on knee joint afferent nerve mechanosensitivity in response to normal (D) and noxious (E) rotation of saline-injected knee joints. The desensitizing effect of ACEA was significantly different from the effect of vehicle control treatment (P < 0.0001 by two-way analysis of variance [n = 10–11 fibers]; ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001, after Bonferroni adjustment). Values are the mean ± SEM.

Effect of selective CB1 and TRPV-1 receptor antagonisms in control knee joints.

In the majority of recorded joint afferent nerves, coadministration of the CB1 receptor antagonist AM251 with ACEA reduced the desensitizing effect of the highest dose of cannabinoid (P < 0.0001 by Student's unpaired t-test; n = 13 fibers) (Figure 2). The inhibitory effect of AM251 was not consistent in all afferent units because ACEA was still able to cause desensitization in 27% of recorded fibers. Similarly, the TRPV-1 receptor antagonist SB366791 also attenuated ACEA-mediated antinociception in control knees (P < 0.0001 by Student's unpaired t-test; n = 10 fibers) (Figure 2). A desensitizing effect of ACEA, however, was still detectable in 11% of the fibers treated with SB366791. Administration of SB366791 alone had no significant effect on the afferent nerve fiber firing rate in the joint (P = 0.90 for non-noxious movement and P = 0.93 for noxious movement, by Student's unpaired t-test; n = 6 fibers) (Figure 2).

Figure 2.

Effect of coadministration of the cannabinoid 1 receptor antagonist AM251 or the transient receptor potential vanilloid 1 receptor antagonist SB366791 with arachidonyl-2-chloroethylamide (ACEA) on control (saline-injected) knee joints. A and B, Effect of AM251 or SB366791 on ACEA-mediated desensitization of knee joint afferent nerve activity with normal (A) and noxious (B) rotation of control joints. AM251 and SB366791 significantly reduced the desensitizing effect of ACEA (∗∗∗ = P < 0.0001 by Student's unpaired t-test). When administered alone, AM251 or SB366791 had no significant effect on joint mechanosensitivity. Values are the mean and SEM of 6–15 fibers.

Effect of ACEA in OA knee joints.

With normal and noxious rotation of OA joints, ACEA caused a decrease in the firing rate in 51% of recorded afferent fibers, 27% of fibers showed no effect of ACEA, while 22% of fibers were hyperresponsive to the drug (i.e., a maximal effect was achieved with the lowest dose of ACEA) (Figure 3A). Application of vehicle had no significant effect on mechanosensitivity (P = 0.49 for non-noxious movement and P = 0.07 for noxious movement, by 1-sample t-test; n = 15–20 fibers). The desensitizing effect of ACEA was maximal 5 minutes after drug application for both types of mechanical stimuli (P < 0.01 by 1-sample t-test; n = 13–15 fibers) and lasted >30 minutes in most fibers. A specimen recording of the inhibitory effect of ACEA on afferent nerve activity in OA joints is shown in Figures 3B and C. The desensitizing effect of ACEA was found to be dose-dependent across the dose range of 10−10 to 10−7 moles, for non-noxious and noxious movements (P < 0.05 by one-way ANOVA). Compared with the vehicle, this desensitizing effect was statistically significant (P < 0.001 by two-way ANOVA) (Figures 3D and E). During normal rotation, the mean ± SEM Emax of ACEA was −57.7 ± 6.92%, and with noxious hyperrotation of the OA knee it was −62.3 ± 7.27%. The Emax of ACEA during noxious movement was significantly greater in OA knees compared with control knees (P < 0.05 by Student's unpaired t-test).

Figure 3.

Effect of arachidonyl-2-chloroethylamide (ACEA) on osteoarthritic (OA) knee joints. A, Proportion of afferent nerve fibers from OA knee joints that showed no change in activity, a dose-dependent decrease in activity, and hyperresponsiveness to local application of ACEA. B and C, Specimen recording of a single unit during normal and noxious rotation of an OA knee before (B) and after (C) close intraarterial application of ACEA. ACEA significantly reduced the firing rate of joint afferent nerve fibers. D and E, Effect of ACEA on knee joint afferent nerve mechanosensitivity in response to normal (D) and noxious (E) rotation of OA knee joints. The desensitizing effect of ACEA was significantly different from the effect of vehicle control treatment (P < 0.001 by two-way analysis of variance [n = 13–20 fibers]; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001, after Bonferroni adjustment). Values are the mean ± SEM.

Effect of selective CB1 and TRPV-1 receptor antagonism in OA knee joints.

The desensitizing effect of ACEA was not affected by coadministration with the CB1 receptor antagonist AM251 (P = 0.59 for non-noxious movement and P = 0.54 for noxious movement, by Student's unpaired t-test; n = 5 fibers) (Figure 4). Coadministration of ACEA with the TRPV-1 receptor antagonist SB366791, however, inhibited the antinociceptive effect of the cannabinoid during normal and noxious rotation (P < 0.0001 by Student's unpaired t-test; n = 9 fibers) (Figure 4). Nevertheless, 25% of recorded fibers in OA knees still responded to ACEA administration. Administered alone, SB366791 had no significant effect on the frequency of afferent nerve fiber firing in OA knees (P = 0.95 for non-noxious movement and P = 0.88 for noxious movement, by Student's unpaired t-test; n = 6–10 fibers) (Figure 4).

Figure 4.

Effect of coadministration of AM251 or SB366791 with ACEA on OA knee joints. A and B, Effect of AM251 or SB366791 on ACEA-mediated desensitization of knee joint afferent nerve activity with normal (A) and noxious (B) rotation of OA joints. AM251 had no effect on ACEA desensitization, while SB366791 significantly reduced ACEA responses in most fibers (∗∗∗ = P < 0.0001 by Student's unpaired t-test). When administered alone, AM251 significantly increased the firing rate of afferent nerve fibers during normal rotation (∗∗ = P < 0.01 by Student's unpaired t-test). Conversely, when administered alone, SB366791 had no significant effect on OA joint mechanosensitivity. Values are the mean and SEM of 5–20 fibers. See Figure 3 for definitions.

Effect of CB1 antagonism on basal firing rate in control and OA joints.

To test the role of endocannabinoids acting via the CB1 receptor in joint mechanosensitivity, the effect of AM251 alone on primary afferent nerve activity was assessed in saline-injected control (n = 13 fibers) and OA (n = 16 fibers) knee joints. Administered alone, AM251 had no significant effect on mechanosensitivity in the control knee joint (Figure 2), but significantly increased the firing rate in all recorded fibers of the OA joint (Figure 4) by up to 102.9 ± 19.01% with non-noxious rotation (P < 0.01 by Student's unpaired t-test). AM251 had no significant effect on firing frequency during noxious rotation of the OA knee, probably due to the fact that units were already firing at a maximum rate in OA joints with noxious rotation (P = 0.12 by Student's unpaired t-test) (Figure 4).

DISCUSSION

In animal models of inflammatory joint disease and OA, it has been shown that joint primary afferent nerves become sensitized (4, 42, 43). Sensitization of these peripheral nerves leads to enhanced mechanosensation in the affected joint, which leads to allodynia, hyperalgesia, and spontaneous pain. Amelioration of peripheral sensitization by physical or pharmacologic means would be beneficial in the management of painful arthritis. Electrophysiologic and behavioral studies provide compelling evidence that cannabinoids can suppress nociceptive transmission through effects at CB1 and CB2 receptors (13, 44, 45). For example, local injection of anandamide into the rat hind paw has been shown to attenuate cutaneous thermal hyperalgesia via a CB1 receptor mechanism (46, 47), while pain behavior after subcutaneous formalin injection was ameliorated by another endocannabinoid, palmitoylethanolamide, acting on CB2 receptors (46). Fox et al provided further evidence of cannabinoid-mediated peripheral analgesia when they showed that a nonselective synthetic cannabinoid agonist was able to block mechanical hyperalgesia in a neuropathic pain model (48).

In the present study, we show for the first time that the selective CB1 receptor agonist ACEA is able to reduce the mechanosensitivity of afferent nerve fibers in control and OA rat knee joints. The desensitizing effect of the cannabinoid was apparent during non-noxious as well as noxious movement of the knee. The antinociceptive effect of ACEA in control joints was blocked by coadministration of AM251, indicating that CB1 receptors are involved in this process. Administration of AM251 by itself had no effect on control joint mechanosensitivity, suggesting that basal endocannabinoid levels are negligible in nonarthritic knees.

The current study also found that the antinociceptive effects of ACEA in rat knees could be blocked by pretreating the animal with the TRPV-1 receptor antagonist SB366791. Precedent for TRPV-1 and cannabinoid interaction comes from studies showing that blockade of TRPV-1 ion channels attenuates cannabinoid-mediated vasomotor control and analgesia (27, 29, 33, 34, 49). In vitro binding studies have clearly shown that SB366791 is highly selective for the TRPV-1 ion channel (50), and in the rat knee, the hyperemic effect of acute capsaicin administration is inhibited by treatment with SB366791 (51). Despite this high selectivity for TRPV-1, it is possible that SB366791 could be blocking peripheral CB1 receptors, although this is unlikely because SB366791 shows poor binding affinity for human CB1 receptors in the central nervous system (50).

A more likely explanation for the inhibitory effect of SB366791 on ACEA-mediated afferent nerve fiber desensitization is that ACEA may be attaching to a unique binding domain on TRPV-1, which leads to deactivation of this pain-sensing cation channel. Evidence of distinct binding sites for other TRPV-1 agonists has previously been reported (52), indicating that TRPV-1 can be differentially modulated by multiple mediators. An interesting observation in these experiments was that following ACEA injection, there was an initial transient burst of activity in the joint sensory units, followed by a more prolonged desensitization phase. This phenomenon is consistent with other TRPV-1 agonists, such as capsaicin and resiniferatoxin (30, 31), and supports the concept of ACEA agonism and subsequent silencing of the TRPV-1 channel.

An alternative explanation is that there may be common biochemical processes that link TRPV-1 and CB1 in peripheral neurons. TRPV-1 sensitivity is related to whether the channel exists in its phosphorylated (sensitized) or dephosphorylated (desensitized) state (53). Phosphorylation of TRPV-1 is controlled by protein kinases whose enzymatic activity is cAMP dependent (53, 54). Since CB1 activation inhibits adenylate cyclase activity, and hence cAMP production (20), ACEA may ultimately cause TRPV-1 dephosphorylation and consequently reduce TRPV-1 sensitivity. Thus, ACEA has the potential to deactivate neuronal TRPV-1 channels either directly (as described above) or by a cAMP-dependent protein kinase dephosphorylation mechanism. The present findings will hopefully provoke future studies to unravel the complex link between CB1 and TRPV-1 modulation of nociceptor function.

Local administration of ACEA to OA knees caused a profound inhibition of joint mechanosensory nerve activity. Of note, the antinociceptive effect of ACEA was greater in the OA joints compared with the saline-injected control knees. Thus, the maximal efficacy of ACEA was significantly higher in OA knees compared with the control knees. In addition, the number of afferent nerve fibers hyperresponsive to ACEA was also noticeably higher in MIA-induced OA knees compared with saline-injected controls. It was found previously that CB1 agonists have a more pronounced effect in inflamed compared with noninflamed tissue (55, 56), and that CB1 expression is increased on primary afferent neurons after induction of inflammation (35). Our finding that induction of OA significantly increases the efficacy of ACEA supports the notion of CB1 receptor up-regulation in this model of degenerative arthritis. A surprising observation in the present series of experiments was that the CB1 receptor antagonist AM251 was unable to block the antinociceptive effects of ACEA in the OA joint. It appears, therefore, that the interaction between ACEA and CB1 receptors in OA knees is so robust that it is difficult to inhibit afferent nerve fiber desensitization with this particular antagonist.

Unlike control joints, local administration of AM251 alone significantly increased mechanosensitivity in the OA knee during non-noxious rotation. This finding suggests that in an OA joint, endocannabinoids are released locally to help offset peripheral sensitization and nociception. The fact that AM251 was ineffective in control knees further corroborates our hypothesis that CB1 receptor expression and sensitivity are enhanced in this model of OA. Another explanation for increased nerve activity following AM251 administration is that the drug may be acting as an inverse agonist in the joint, leading to peripheral sensitization; however, this possibility is unlikely since no increase in afferent nerve activity was observed after AM251 administration in the normal knee joint. In contrast to the present findings, recent data have shown that AM251 produces heat hyperalgesia in normal but not in inflamed tissue, indicating a down-regulation of the endogenous cannabinoid system during inflammation (35). These conflicting results might be due to differences in the inflammatory model used, the route of drug application, or the method of pain assessment.

In this study it was found that local administration of the CB1-selective agonist ACEA significantly suppresses joint nociceptive transmission by reducing primary afferent nerve activity. Blockade of either the CB1 receptor or the TRPV-1 channel attenuated ACEA responses, indicating that both pathways are involved in ACEA-mediated antinociception. The antinociceptive effect of ACEA was enhanced in the OA joint, suggesting an up-regulation of cannabinoid activity in this model of degenerative joint disease. Thus, modulation of the cannabinoid and vanilloid systems could be a useful approach to suppress OA pain.

AUTHOR CONTRIBUTIONS

Dr. McDougall had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Schuelert, McDougall.

Acquisition of data. Schuelert.

Analysis and interpretation of data. Schuelert, McDougall.

Manuscript preparation. Schuelert, McDougall.

Statistical analysis. Schuelert.

Acknowledgements

We thank Benjamin Andruski and Chunfen Zhang for their technical assistance in this project.

Ancillary